ture. The feasibility of estimating district level in- season rice crop area using two and three-date ERS-
Ž SAR data has been demonstrated Panigrahy et al.,
.
1
1997 . However, the results obtained for tuber crops Ž
. Ž
including potato were not very encouraging ESA, .
1995 . The scope of possible utilisation of SAR data for
large-area crop survey has widened with the launch of the Canadian satellite, RADARSAT in November
1995. RADARSAT is designed specifically to pro- vide advanced SAR imaging capability. It operates at
a C-band frequency of 5.3 GHz, with horizontal polarisation for both transmission and reception of
the signals and can be steered to image a swath from 50 to 500 km using a number of imaging modes.
There are seven beams with 25 beam positions pro- viding data with a range of incidence angles from 10
to 59 degrees and spatial resolutions from 10 to 100 m. The Standard beam data have seven beam posi-
tions with incidence angle varying from 208–498 Ž
. Anon, 1996 . Shallow angle data of more than 408
incidence angle is of particular interest, as it is considered better suited for identification of agricul-
tural crops. SAR data has a significant role in the remote
sensing based crop survey programme in India due Ž
to the cloud cover problem during monsoon July to .
October , the main crop season. This paper describes the results of the investigations carried out using
RADARSAT Standard Beam 7 data for rice and potato crops in their predominant growing environ-
ments in West Bengal, India. The work has been carried out under the project entitled ‘‘RADARSAT
data evaluation for crop identification and characteri- zation’’, under the Application DeÕelopment and
Ž Research Opportunity ADRO programme ADRO
. project ID 349
of RADARSAT, sponsored by Ž
. RADARSAT International
RSI and the Canada
Ž .
Centre of Remote Sensing CCRS , Canada.
2. Study area
The study area lies in the Bardhaman and Hooghli districts of West Bengal state in Eastern India lo-
1
Tuber crops: crops with a tuber, an enlarged, fleshy, under- ground stem with buds capable of producing new plants.
cated between 23800
X
to 23830
X
N latitude and 87840
X
to 88820
X
E longitude. Potato crops, grown during the Ž
. rabi season December–March , and rice crop, grown
Ž .
during the summer season February–May , are ad- dressed in this study. Around 70 of potato fields
were sown in the first week of December and har- vested by end of February. The rest were sown at the
end of December and harvested in March. These two categories were designated as early and late potato,
respectively. Rice was grown totally under wetland irrigated practice in summer. Seedlings were trans-
planted in puddled fields filled with water during end of January to middle of February. The crop was
harvested in May.
3. Data used
RADARSAT S7 data have been used in this study. The incidence angle of beam S7 data ranges
from 458–498, covering a nominal area of 100 = 100 km
2
. The data product used belonged to the ‘‘Path Image Plus’’, which is 16-bit data having a pixel size
2 2
Ž of 8
m . Four-date descending-node data equa- .
torial pass time 0630 h acquired on January 2 and 26, February 19 and March 15, 1997 were used. The
crop growth stages of rice and potato in relation to the SAR data acquisition are given in Table 1.
Optical remote sensing data from the Indian Remote
Ž .
Sensing satellite IRS 1C, i.e., LISS-III data ac- quired on January 26, 1996, October 1996 and PAN
data acquired on February 12, 1996, were used to Ž
. generate base maps for ground truth GT data col-
Ž lection. LISS-III multispectral data green, red and
. near infrared bands
have a spatial resolution of
Table 1 Crop stages in relation to RADARSAT data acquisition dates
Ž .
Date Potato earlyrlate
Rice 2.1.97
Vegetative Fallow fields
26.1.97 Tuberingrvegetative
Fallowrpuddling
a
19.2.97 Maturertubering
Puddlingrearly tillering 15.3.97
Fallowrmature Peak tilleringrpanicle
a
Tillering: growth stage for cereal crops, when additional shoots are developing from the crown.
23 m and the panchromatic data has a spatial resolu- tion of 5.8 m.
4. Methodology
